Computational Design Optimization of All-Solid-State Lithium Ion Battery Electrode by Multi-Scale Simulation Based on Finite Element Method Combined with Density Functional Theory

Abstract

One of the major challenges in all-solid-state lithium-ion battery (ASSLiB) developments for electronic vehicles is to prevent crack formation in electrodes during charge and discharge. The crack forms due to the volumetric expansion caused by lithium ion insertion to active materials. Although X-ray or electron beam based analysis can provide, the cracks in the electrodes can be analyzed using limited and heavy cost analysis, that can be a hurdle to optimize durable battery design. In order to accelerate the structural analysis and the design optimization, a simulation technology to predict the structural changes has been highly desired. The electrode degradation involves complicated phenomena such as powder deformations, volumetric expansion of active materials caused by the lithium insertion, interface peelings on the powder interfaces and the crack formation. This complexity has been the main reason why the practical simulation methodology has not emerged so far. To address the issue, we propose a new simulation methodology based on a finite element method (FEM) combined with a density functional theory (DFT). In the electrode degradation simulation, we take a new FEM approach called multi-particle FEM (MP-FEM). Since many FEM work treat compacted powders as a continuum body, they cannot describe the peelings at the particle interfaces or cracks in the electrode. On the other hand, MP-FEM can describe these phenomena since MP-FEM treats each particle structures explicitly. We apply a tiebreak model to the particle interfaces for accurate simulation. To perform MP-FEM, material properties must be set properly. In evaluating the material properties, we recommend to apply DFT simulation at the same time with experimental evaluations. Normally, material properties are evaluated by tests. However, the precise evaluation is challenging and time consuming because of several reasons such that the evaluation on normal test piece cannot be applied since the target materials are powders with the diameters around 0.1~100 micro meter. To demonstrate the advantage of our methodology, we built an electrode model containing an active material (AM) and a solid electrolyte (SE). In the first step of the simulation, the model electrode was pressed. After the removal of the all pressure, pore volume ratio was predicted as 20%, which showed a good agreement with experimental value of 18%. In the second step, the AM powders were controlled to expand and shrink representing charge and discharge. Our model reproduced the brake down of connected powder interfaces indicating decrease of lithium conduction paths. The volumetric ratio of pores in the electrode decreased during the charge process which also showed a good agreement with an experimental finding that internal resistance of the electrode decreased during charge processes. Based on these results, we believe that the methodology can predict electrode durability. In this paper, we will also compare several electrodes with varied design parameters and discuss design guidelines for durable electrode design.